In-Built Customised Mechanical Failure of 316L Components ... - MDPI

technologies Article

In-Built Customised Mechanical Failure of 316L Components Fabricated Using Selective Laser Melting Andrei Ilie, Haider Ali and Kamran Mumtaz * Centre for Advanced Additive Manufacturing, Department of Mechanical Engineering, University of Sheffield, Sheffield S1 3JD, UK; [email protected] (A.I.); [email protected] (H.A.) * Correspondence: [email protected]; Tel.: +44-114-222-7789 Academic Editors: Salvatore Brischetto, Paolo Maggiore and Carlo Giovanni Ferro Received: 31 January 2017; Accepted: 21 February 2017; Published: 25 February 2017

Abstract: The layer-by-layer building methodology used within the powder bed process of Selective Laser Melting facilitates control over the degree of melting achieved at every layer. This control can be used to manipulate levels of porosity within each layer, effecting resultant mechanical properties. If specifically controlled, it has the potential to enable customisation of mechanical properties or design of in-built locations of mechanical fracture through strategic void placement across a component, enabling accurate location specific predictions of mechanical failure for fail-safe applications. This investigation examined the process parameter effects on porosity formation and mechanical properties of 316L samples whilst maintaining a constant laser energy density without manipulation of sample geometry. In order to understand the effects of customisation on mechanical properties, samples were manufactured with in-built porosity of up to 3% spanning across ~1.7% of a samples’ cross-section using a specially developed set of “hybrid” processing parameters. Through strategic placement of porous sections within samples, exact fracture location could be predicted. When mechanically loaded, these customised samples exhibited only ~2% reduction in yield strength compared to samples processed using single set parameters. As expected, microscopic analysis revealed that mechanical performance was closely tied to porosity variations in samples, with little or no variation in microstructure observed through parameter variation. The results indicate that there is potential to use SLM for customising mechanical performance over the cross-section of a component. Keywords: additive manufacturing; mechanical properties; customisation; selective laser melting

1. Introduction and Background Research in the field of Selective Laser Melting (SLM) is broad, with many focused on parameter optimisation for achieving consistent high density parts and establishing a relationship between parameters and final part mechanical properties for specific materials [1–13]. Understanding the phenomenon of residual stresses building up in SLM parts due to high thermal gradients and strategies for minimising the residual stresses and their detrimental effects such as cracking and warping of parts are other areas of interest for researchers [14–21]. Analysis of post-processing operations such as hipping and heat treatment of additively manufactured parts has also been a point of interest for researchers as these can enable significant improvement in part density and mechanical properties [15,18,22,23]. Despite the increased interest and growth in SLM research, the complexity of this rapid solidification process has resulted in high value industries approaching this technology with caution [24]. In addition, the requirements for process repeatability, tolerances, and feedstock traceability have increased in recent years [25] and are understandably strict in sectors such as aerospace and biomedical orthopedics. Technologies 2017, 5, 9; doi:10.3390/technologies5010009

www.mdpi.com/journal/technologies

Technologies 2017, 5, 9

2 of 12

SLM of stainless steel alloys, particularly 316L, has been of major interest to researchers due to its high corrosion resistance, formability, strength, weldability, and biocompatibility. SS 316L is suitable for medical applications (implants and prosthesis), pharmaceuticals, architectural applications, fasteners, aerospace parts, marine and chemical applications, and heat exchangers [26,27]. SS 316L SLM parts in as-built conditions have exhibited fatigue properties similar to conventionally manufactured parts, mainly due to the high ductility even after SLM processing [7]. If the correct combination of process parameters are chosen, SLM steel samples usually exhibit improved tensile properties compared to conventionally manufactured steel samples [28]. Customised Mechanical Properties Using SLM The layer by layer building methodology used within SLM offers the potential for mechanical properties to be controlled layer to layer. However, to date, limited research has been undertaken exploring this possibility. Theoretically this variation in mechanical properties can be generated through a number of methods. The first and most obvious is by geometry manipulation, designing with the assistance of computational software can allow designers to create structures that will have a specific mechanical response to external loading. This feature is not exclusive to additive manufacturing technologies and can be achieved through a variety of other manufacturing processes (however, complexity may be limited when using conventional processes). Secondly properties can be controlled through the introduction of variable materials (i.e., functionally graded materials), although this is challenging due to the potential for thermal expansion mismatch between materials leading to delamination of multi-material layers, it is also difficult to recycle and separate materials from a powder bed that consists of graded multi-materials. Thirdly, microstructure manipulation could be used to vary mechanical properties across the cross-section of parts, this may be achieved through adjustment of melting regimes employed at each layer [3,29]. However, due to the rapid solidification of material within SLM, it is often difficult to alter the microstructure of individual layers significantly from their standard fine dendritic form solely through parameter control (i.e., laser power, exposure time, etc.). Suitable SLM processing conditions (i.e., promoting high density components) and rapid solidification limit generation of sufficient variability within the thermal history of each layer to promote enough microstructural change to allow mechanical performance to vary significantly. Microstructures can be altered through multiple reheating strategies and powder bed pre-heating, but this tends to affect multiple layers across a component and precise layer to layer control will be challenging. Pre-heating the powder bed to higher temperatures can assist in delaying solidification and coarsen the microstructure, however this heating will affect the majority of layers across a component without permitting control over microstructural variation across specific layers. Finally, mechanical properties could be altered through the introduction of controllable features such as porosity across a component. This can be controlled through varying laser processing parameters layer to layer and inducing lack of fusion porosity. The mechanical properties can be artificially altered, making the areas across a component mechanically weaker in the tensile compared to a fully dense region, this can be designed across a part, promoting preferential deformation of parts or failure at strategic locations. Currently, when tensile testing SLM components the exact location of failure is random and highly unpredictable without the use of X-ray analysis. Customising the mechanical performance of an SLM sample may be particularly useful for applications requiring a fail-safe mechanism to minimise harm or respond appropriately to specific types of loading conditions, examples include pressure relief in pressure valves or blow-off panels used in enclosure. Designed premature failure may also be used to prevent more expensive or catastrophic failure from occurring further down the line. 2. Experimental Methodology As an overview of the experimental methodology, processing parameters were adjusted so that variation in sample density could be attained. Having established what effect these adjusted

Technologies 2017, 5, 9

3 of 12

2. Experimental Methodology

Technologies 2017, 5, 9

3 of 12

As an overview of the experimental methodology, processing parameters were adjusted so that variation in sample density could be attained. Having established what effect these adjusted parameters parameters had, had, samples samples were were created created such such that that variation variation in in mechanical mechanical properties properties (i.e., (i.e., preferential preferential point of failure) would be created across strategic locations within samples. Parts point of failure) would be created across strategic locations within samples. Parts were were tested tested for for density, hardness, and tensile strength. density, hardness, and tensile strength. A A Renishaw Renishaw SLM SLM 125 125 was was used used during during investigations, investigations, this this system system uses uses aa 200 200 W W fibre fibre laser laser to to process metallic powdered feedstock within a purged argon atmosphere. Gas atomised SS AISI 316L process metallic powdered feedstock within a purged argon atmosphere. Gas atomised SS AISI 316L (15–45 (15–45 micron) micron) was was used used as as the the feedstock feedstockmaterial, material,its itscomposition compositionisisshown shownin inTable Table1.1. Table 1. Weight % Composition. Table 1. Weight % Composition. Element Element Fe

%Composition %Composition Bal

FeC C Si Bal 0.012 0.012 0.6

S S Cr Cr NiNi Si MnMn P P 0.61.251.25 0.012 0.012 0.005 0.005 17.817.8 12.9 12.9

Mo Mo

2.35 2.35

N N 0.04 0.04

Cu Cu OO 0.03 0.0185 0.03 0.0185

2.1. Sample Sample Testing Testing 2.1. 10 × × 10 hardness testing testing machine machine 10 10mm mmcubes cubeswere were fabricated fabricated for for testing testing hardness hardness using using aa Vickers Vickers hardness (BSEN ISO 6507-1:2005 [30]). The level of porosity in the specimen was estimated by area fraction (BSEN ISO 6507-1:2005 [30]). The level of porosity in the specimen was estimated by area fraction analysis of representative micrographs/fields using a method based on ASTM E2109-01 (2007) analysis of representative micrographs/fields using a method based on ASTM E2109-01 (2007) andand BS BS 7590:1992 [31,32]. Tensile test specimens were manufactured according to the ASTM E8 standard 7590:1992 [31,32]. Tensile test specimens were manufactured according to the ASTM E8 standard [33]. [33].specimens The specimens were on tested on aOlsen Tinius OlsenUTM H25K-S UTMMaterials BenchtopTester. Materials Tester. of A The were tested a Tinius H25K-S Benchtop A summary summary of keyofdimensions of the tensile pieces are shown in samples Figure 1.were The built samples were built key dimensions the tensile test pieces aretest shown in Figure 1. The vertically and vertically and tested along this axis. tested along this axis.

Figure Figure1.1. Specimen Specimen dimensions dimensions(a); (a);and andCAD CADfile file(b). (b)

2.2. Processing Parameter Parameter Selection Selection 2.2. For processing processing SS SS 316L 316L aa set set of of recommended recommended laser laser process process parameters parameters were were specified specified by by SLM SLM For OEM Renishaw Renishaw for for creating creating components components at atfull full density, density,shown shownin inTable Table2.2. OEM Table 2. 2. Manufacturer Manufacturer (Renishaw) (Renishaw) recommended recommended processing processingparameter parametervalues valuesfor forSS SS316L. 316L. Table

Parameter Parameter

Value Value

Layer Thickness—L (µm) Layer Thickness—L (µm) Point Distance—x(µm) (µm) Point Distance—x Hatch Spacing—h(µm) (µm) Hatch Spacing—h Spot SpotSize Size(µm) (µm) Laser Power—P Laser Power—P(W) (W) Exposure time—t (s) Exposure time—t (s)

50 µm 50 µm 50 µm 50 µm 90 µm µm 50 µm 50 µm 200 200 W W 70 µs 70 µs

Various processing parameters affect the SLM process, these may be direct or indirect as specified by Yadroitsev et al. [2]. However, the principal parameters in SLM—those which have the most substantial effect—have been identified as being laser power, spot size diameter, exposure time,

Technologies 2017, 5, 9 Technologies 2017, 5, 9 Technologies 2017, 5, 9

4 of 12 4 of 12 4 of 12

Various processing parameters affect the SLM process, these may be direct or indirect as Various processingetparameters affect the theprincipal SLM process, these inmay be directwhich or indirect as specified by Yadroitsev al. [2]. However, parameters SLM—those have the specified by Yadroitsev et al. [2]. However, the principal parameters in SLM—those which have the mostspacing substantial effect—have been identified aspowder being laser power, spot [1]. sizeVarying diameter, exposure spot (scan speed), hatch distance, and layer thickness laser powertime, and most substantial effect—have been identified aspowder being laser power, spot sizeVarying diameter, exposure time, spot spacing (scan speed), hatch distance, and thickness laser power and scan speed individually affects the laser energy densitylayer that is used to [1]. melt the powder bed during spot spacing (scan speed), hatchthe distance, and powder layer thickness [1]. Varying laser bed power and scan speed individually affects laser energy density that is used to melt the powder during SLM, and controls levels of porosity within a component. This implies that there is an “optimum” scan speed individually affects the laser energy density that is used to melt the powder bed during SLM, and controls of porosity withinparts a component. This implies there is an “optimum” energy density for levels fabricating fully dense free of irregularities as that noted by Kurian et al. [10]. SLM, and controls levels of porosity within a component. This implies that there is an “optimum” energy density fabricating fully is dense parts freeEquation of irregularities aswas noted by Kurian et al. [10]. The laser energyfor density behaviour described by (1), and it shown that laser power energy density for fabricating fully is dense parts by freeEquation of irregularities as noted by Kurian et al. [10]. The laser energy density behaviour described (1), and it was shown that laser power and exposure time are inversely proportional (other variables kept constant), in theory it is possible The laser energy density behaviour is described by Equation (1), and it was shown thatitlaser power and exposure time are inversely proportional (other kept constant), in theory is possible to vary both the laser power and exposure time in variables unison while maintaining the energy density and exposure time are inversely proportional (other variables kept constant), in theory it is possible to vary both the laser power and exposure time in unison while maintaining the energy density delivered constant. to vary both the laser power and exposure time in unison while maintaining the energy density delivered constant. Pt delivered constant. Energy density = Q = (1) = =xhl (1) = =distance (mm); l: Layer Thickness (mm); (1) P: Laser power (W, J/s); t: Exposure time (s); x: Point 3 P: Laser power (W, J/s); t: Exposure time (s); x: Point distance (mm); l: Layer Thickness (mm); h: h: Hatch Spacing (mm); Q: Energy density (J/mm ). P: Laser power (W, J/s); t: Exposure time (s); x: Point distance (mm); l: Layer Thickness (mm); h: ). Hatch Spacing (mm); Q: Energy density For these reasons, a variation of laser (J/mm power with exposure time was performed, while maintaining ). Hatch Spacing (mm); Q: Energy density (J/mm 3 , consistentwhile For these reasons, a variation laser power with selected exposure time was performed, an arbitrary “optimal” energy density.ofThe energy density was 62 J/mm with For these reasons, a variation of laser power withenergy exposure time selected was performed, while, maintaining an arbitrary “optimal” energy density. The density was 62 J/mm the Renishaw SLM 125 advised processing parameters. This energy density was maintained in maintaining an the arbitrary “optimal” energy density. The energy density selected wasdensity 62 J/mm consistent with Renishaw SLM 125 advised processing parameters. was, combination with varying exposure times and laser powers as shown in This Tableenergy 3, continuous melt consistent withcombination the Renishaw SLM 125 advised processing parameters. Thisasenergy density was maintained with varying laser shown in Table 3, tracks were in identified at powers as low asexposure 150 W intimes other and work [9]. powers Other SLM parameters were maintained in combination with varying exposure times and laser powers as shown in Table 3, continuous melt tracks were identified at powers as low as 150 W in other work [9]. Other SLM kept constant. continuous melt tracks were identified at powers as low as 150 W in other work [9]. Other SLM parameters were kept constant. parameters were kept constant. Table 3. Varied parameter range and processing set name/condition. Table 3. Varied parameter range and processing set name/condition. Table 3. Varied parameter range and processing set name/condition. Specimen Set A B C

Specimen Set

Specimen Set Laser power (W) 200 Laser power (W) Exposure time (µs) 70 Laser power (W)

Exposure time (µs) Exposure time (µs)

A A 200 200 70 70

B C B 175 C 175 150 80 175 80 150 93 80 93

150 93

Using Using the the Renishaw Renishaw SLM SLM 125 125 and and processing processing SS SS 316L 316L powder, powder, three three repeat repeat samples samples of of each each Using the Renishaw SLM 125 and processing SS 316L powder, three repeat samples of each specimen type A, B, and C were produced as shown in Figure 2. specimen type A, B, and C were produced as shown in Figure 2. specimen type A, B, and C were produced as shown in Figure 2.

Figure 2. Uniform specimen sets. Figure Figure 2. 2. Uniform Uniform specimen specimen sets. sets.

The specimens were attached to the substrate plate via support structures in order to facilitate facilitate The were attached to the substrate plate support structures order to facilitate removal. specimens Images of the support structures, including thevia surface texture of the in specimen, with top removal. Images of the support structures, including the surface texture of the specimen, with top removal. Images of the support structures, including the surface texture of the specimen, with top and and bottom (chiseled) sides are shown in Figure 3. and bottom (chiseled) are shown in Figure bottom (chiseled) sidessides are shown in Figure 3. 3.

Figure 3. Support structures (a); and top (b); bottom (c) surface texture. Figure 3. 3. Support structures (a); and top (b); bottom (c) surface texture.

Technologies 2017, 5, 9

5 of 12

Technologies 2017, 5, 9 3. Results

5 of 12

Technologies 2017, 5, 9

5 of 12

Three sets of SS 316L cubed samples and tensile test bars were produced for each processing 3. Results Technologies 2017, 5, 9 5 of 12 3. Results condition (A, B, C) using SLM parameters shown in Table 3. Based on these results, further customised Three sets of SS 316L cubed samples and tensile test bars were produced for each processing samples were and tested, their results are laterin detailed in Based Section 3. Results Threecreated setsB,ofC) SS 316L cubed samples and tensile testTable bars 3. were produced for each processing condition (A, using SLM parameters shown on3.3. these results, further condition (A, B, C) using SLM parameters shown in Table 3. Based on these results, further customised samples were created and tested, their results are later detailed in Section 3.3. Three sets of SS 316L cubed samples and tensile test bars were produced for each processing 3.1. Optical Microscopy customised samples were created and tested, their results are later detailed in Section 3.3. condition (A, B, C) using SLM parameters shown in Table 3. Based on these results, further 3.1. Optical Microscopy Microscopy was performed as described Section 2.1,are initially testingincubed samples for porosity customised samples were created and tested,intheir results later detailed Section 3.3. 3.1. Optical Microscopy Microscopy performed as described Section 2.1,setinitially cubeddensity samplesoffor and shown in Figurewas 4. Samples produced withinparameter A hadtesting an average 99.8%, 3.1. Optical Microscopy was performed as described in Section 2.1, initially for porosity shown in Figure 4. Samples with set Atesting had an cubed average density of parameter setand BMicroscopy 99.1%, and parameter set produced C 96.8%. Evenparameter though energy density wassamples maintained at 3 , as porosity and inperformed 4. reduced Samples produced with parameter setthough Atesting had an average density of 99.8%, parameter set BFigure 99.1%, and parameter C 96.8%. Even energy density was 62 J/mm the shown laser power was the porosity within samples increased due to a potential Microscopy was as described insetSection 2.1, initially cubed samples for 99.8%, parameter set B andpower parameter setwith C 96.8%. Even though energy density was maintained atshown 62 J/mm , as99.1%, the4.laser was reduced the porosity within increased due porosity and inevident Figure Samples produced parameter set A hadsamples an average density of increase in lack of fusion, from the irregularly shaped pores. Subsequently, the etched samples maintained at 62 J/mm , as the laser power was reduced the porosity within samples increased due to a potential increase in lack of fusion, evident from the irregularly shaped pores. Subsequently, the parameter set detailed B 99.1%, view and of parameter set C 96.8%. Even inthough density was were 99.8%, examined for a more the microstructure shown Figureenergy 5, microstructures were to a potential increase in, lack oflaser fusion, evident from the irregularly shaped pores. Subsequently, the etched samples were examined for apower more detailed view of the microstructure shown in Figuredue 5, maintained at 62 J/mm as the was reduced the porosity within samples increased typical of those produced using SLM little or no variation in microstructure between samples etched samples were forthose awith more detailed view of SLM the microstructure in Figure 5, microstructures wereexamined produced using with little orshown no variation in to a potential increase intypical lack of of fusion, evident from the irregularly shaped pores. Subsequently, the whenmicrostructure melting at different laser powers. microstructures were typical of those produced using SLM with little or no variation in between samples when melting at different laser powers. etched samples were examined for a more detailed view of the microstructure shown in Figure 5, microstructure between samplesofwhen melting at different powers. microstructures were typical those produced usinglaser SLM with little or no variation in microstructure between samples when melting at different laser powers.

(A) (B) (C) (A) (B) (C) Figure 4. Porosity of cubes fabricated using parameters (A) (200 W); (B) (175 W); (C) (150 W). Figure 4. Porosity of cubes fabricated using parameters (A) (200 W); (B) (175 W); (A) (B) (C)(C) (150 W). Figure 4. Porosity of cubes fabricated using parameters (A) (200 W); (B) (175 W); (C) (150 W).

Figure 4. Porosity of cubes fabricated using parameters (A) (200 W); (B) (175 W); (C) (150 W).

(A) (A)

(B) (B)

(C) (C)

Figure 5. Microstructure of cubes fabricated using parameters (A) (200 W); (B) (175 W); (C) (150 W). (A) (B)parameters (A) (200 W); (B) (175 W); (C)(C) (150 W). Figure 5. Microstructure of cubes fabricated using

Figure 5. Microstructure of cubes fabricated using parameters (A) (200 W); (B) (175 W); (C) (150 W). 3.2. Tensile Testing Figure 5. Microstructure of cubes fabricated using parameters (A) (200 W); (B) (175 W); (C) (150 W). 3.2. Tensile Testing Tensile testing was performed on each specimen set (A, B, and C). The specimens fractured at 3.2. Tensile Testing 3.2. Tensile Testing Tensile testing performed onaseach (A, B, C).beThe fractured at varying points alongwas the gauge length can specimen be seen inset Figure 6. and It can seenspecimens that fracture location Tensile testing was performed on each specimen set (A, B, and C). The specimens fractured at varying points along the gauge length as can be seen in Figure 6. It can be seen that fracture location across the gauge length of samples are random and unpredictable. Tensile testing was performed on each specimen set (A, B, and C). The specimens fractured at across the gauge length ofgauge samples are random varying points along thethe gauge length asas can seen in Figure Figure6.6.It Itcan can seen fracture location varying points along length canbe beand seenunpredictable. in bebe seen thatthat fracture location

acrossacross the gauge length of samples are the gauge length of samples arerandom randomand and unpredictable. unpredictable.

Figure 6. Set B (175 W) specimen fracture locations. Figure 6. Set B (175 W) specimen fracture locations. Figure 6. Set B (175 W) specimen fracture locations.

Figure 6. Set B (175 W) specimen fracture locations.

Technologies 2017, 5, 9

6 of 12

Technologies 2017, 5, 9 Figure 7 shows

of 12 stress-strain plots from tensile testing for one representative sample from6each processing condition. at Technologies 2017, 5, 9 It reveals that a 3% porosity variation between samples A–C processed 6 of 12 constant energy densities (but varying laser powers) was sufficient to generate variation in Figure 7 shows stress-strain plots from tensile testing for one representative sample from each Figure 7 shows As stress-strain fromgenerally tensile testing for one representative sample yield from each mechanical properties. expected,plots trends indicate weaker parts (reduced strength, processing condition. It reveals that aa3% porosity variationbetween between samples A–C processed at processing condition. It reveals that 3% porosity variation samples A–C processed at part UTS, and fracture strength) were formed with use of lower power lasers due to higher levels of constant energy densities (but varying laser powers) was sufficient to generate variation in mechanical constant energy densities varying laserrepeats) powers) sufficient to4 generate variation in porosity. The mean average for(but all samples (×3 is was shown in Table and Figure 8. properties. As expected, generally indicate weaker parts (reduced yield strength, UTS, and mechanical properties.trends As expected, trends generally indicate weaker parts (reduced yield strength, UTS, and fracture strength) formed usepower of lowerlasers powerdue lasers due to higher of part fracture strength) were formedwere with use ofwith lower to higher levelslevels of part porosity. porosity. The mean average for all samples (×3 repeats) is shown in Table 4 and Figure 8. The mean average for all samples (×3 repeats) is shown in Table 4 and Figure 8.

Figure 7. Uniform specimen stress-strain plot. Figure Uniformspecimen specimen stress-strain Figure 7.7.Uniform stress-strainplot. plot. Table 4. Mean mechanical property valuesfor for uniform specimen sets. Table 4. Mean mechanicalproperty property values values sets. Table 4. Mean mechanical foruniform uniformspecimen specimen sets.

Specimen Specimen

A (200 W) BB(175 (175W) W) C C (150 (150 W)W)

A (200 W) Property Property Specimen A (200 W) Property 0.2% Yield point (MPa) 443 ± 11 0.2% Yield point (MPa) 443 0.2% Yield point (MPa) 443 ±±117 UTS (MPa) 565 565 UTSUTS (MPa)(MPa) 565 ± ±1717 Fracture stress(MPa) (MPa) 558 558 ±1313 Fracture stressstress (MPa) ±± Fracture 558 13

B (175 W)

C (150 W)

385 344344 ± 5± 5 385± ±3 3 385 ± 3 425 ± 22344 ± 5 483 ± 9 483 425 ±425 22± 22 483±±99 473 ± ±10 ± 33 473 ±10 10 405 473 405 ±405 33± 33

Figure 8. Comparison of mean mechanical properties between samples (A) (200 W); (B) (175 W); (C) (150 W).

Figure 8. Comparison properties between samples (A)(A) (200 W);W); (B) (B) (175(175 W); W); (C) Comparisonofofmean meanmechanical mechanical properties between samples (200 (150 W). (C) (150 W).

Technologies 2017, 5, 9

7 of 12

Technologies 2017, 5, 9 3.3. Hardness Testing Results

7 of 12

HardnessTesting tests were performed on cubic SLM samples. The mean average result for each sample 3.3. Hardness Results is shown in Table 5. The higher apparent scan speeds (generated with shorter exposure times) used Hardness tests were performed on cubic SLM samples. The mean average result for each sample in sample A led to increased sample hardness due to faster solidification of the melt pool. Slower is shown in Table 5. The higher apparent scan speeds (generated with shorter exposure times) used apparent scan speeds (as those used in sample C) lead to slower solidification rates and therefore a in sample A led to increased sample hardness due to faster solidification of the melt pool. Slower softer material. Higher hardness values generally indicate higher tensile strength and are therefore in apparent scan speeds (as those used in sample C) lead to slower solidification rates and therefore a agreement with the tensile results for each sample detailed in Section 3.1. softer material. Higher hardness values generally indicate higher tensile strength and are therefore in agreement with the tensile results for each sample detailed in Section Table 5. Uniform specimen hardness testing results.3.1. Table 5. Uniform specimenAhardness testingBresults. Sample Average Vickers Hardness (HV) Sample

Average Vickers Hardness (HV)

193 ± 1 A

193 ± 1

C

159 ± 4C B 185 ± 3 159 ± 4

185 ± 3

3.4. Customised Specimens 3.4. Customised Specimens The results obtained for the uniform sample set show that there is significant variance in the The results obtained for the uniform sample set show that there is significant variance in the mechanical properties as energy density remains constant and laser powers and exposure times mechanical properties as energy density remains constant and laser powers and exposure times are are adjusted. The uniform sample set fractured in unpredictable locations within the gauge length adjusted. The uniform sample set fractured in unpredictable locations within the gauge length (Figure 6). In order to customise mechanical properties and create predictable failure/break locations (Figure 6). In order to customise mechanical properties and create predictable failure/break locations across specimens, two different sets of processing parameters were used to build a single tensile across specimens, two different sets of processing parameters were used to build a single tensile test test specimen to initiate “structural change” within the sample. “Failure points” were introduced specimen to initiate “structural change” within the sample. “Failure points” were introduced in the in the customised specimens at different locations along the gauge length of each sample as shown customised specimens at different locations along the gauge length of each sample as shown in in Figure 9. Figure 9.

Figure 9. Designed failure locations for customised specimen set. Figure 9. Designed failure locations for customised specimen set.

In Inorder orderfor forthe thecustomised customisedspecimens specimensto tobe beproduced producedon onthe theSLM SLMmachine, machine,each eachspecimen specimenwas was split into three parts; the bottom end, the top end, and the failure point. This distinction wasmade made split into three parts; the bottom end, the top end, and the failure point. This distinction was in in order for the failure point to be assigned a different processing parameter set compared to the order for the failure point to be assigned a different processing parameter set compared to the top/bottom For the the customised specimen sets sets D, top/bottom ends. ends. For customised specimen D, E, E, and and FF (Figure (Figure 9), 9), the the failure failurepoints pointswere were placed 5 mm respectively from thethe toptop section of the length. The failure pointpoint was placedatat15, 15,10, 10,and and 5 mm respectively from section of gauge the gauge length. The failure designed to be 10 layers thick (i.e., 500 µm) to encourage an adequate or distinct change in mechanical was designed to be 10 layers thick (i.e., 500 µm) to encourage an adequate or distinct change in performance. The processing 200 W laser µs exposure for the mechanical performance. Theparameters processing were parameters were power 200 W with laser 70 power with 70 time µs exposure top/bottom of the sample parameter setparameter A), and 150 with µs W forwith the failure point time for theends top/bottom ends of(original the sample (original setW A), and93150 93 µs for the (original parameter set C). These parameters were selected from the uniform sample sets A and C as failure point (original parameter set C). These parameters were selected from the uniform sample they displayed the largest variation in mechanical properties as detailed in Section 3.2. The specimens sets A and C as they displayed the largest variation in mechanical properties as detailed in

Section 3.2. The specimens produced are geometrically identical to the uniform tensile samples produced in Section 3.2, however, the failure point locations were labelled on the customised specimens for better visualisation, as shown in Figure 10.

Technologies 2017, 5, 9

8 of 12

produced are geometrically identical to the uniform tensile samples produced in Section 3.2, however, the failure point locations were labelled on the customised specimens for better visualisation, as shown Technologies 2017, 5, 9 8 of 12 in Figure 10. Technologies 2017, 5, 9

8 of 12

Figure 10. Customised Customised specimen set failure points. Figure 10. Customised specimen set failure points.

Tensile tests were were performed on on the customised specimen sets D, E, Tensile E, and and F. F. All All samples samples fractured fractured Tensile tests tests were performed performed on the the customised customised specimen specimen sets sets D, D, E, and F. All samples fractured at the marked locations for designed in-built failure point, shown in Figure 11. This is evidence that at the marked locations for designed in-built failure point, shown in Figure 11. This is evidence at the marked locations for designed in-built failure point, shown in Figure 11. This is evidence that that predictable and controlled controlled fracture/failure fracture/failure can be achieved through selection and implementation of predictable and can be achieved through selection and implementation predictable and controlled fracture/failure can be achieved through selection and implementation of of multiple “hybrid” laser processing parameters within a single build. multiple multiple “hybrid” “hybrid” laser laser processing processing parameters parameters within within aa single single build. build.

(D) (D)

(E) (E)

(F) (F)

Figure 11. Customised specimen controlled fracture locations (D–F). Figure fracture locations locations (D–F). (D–F). Figure 11. 11. Customised Customised specimen specimen controlled controlled fracture

The mechanical properties of the sample sets D, E, and F are shown in Table 6 with the stressThe mechanical properties of the sample sets D, E, and F are shown in Table 6 with the stressstrainThe plot for these properties samples shown in Figure expected, the mechanical properties of these mechanical of the sample sets12. D, As E, and F are shown in Table 6 with the stress-strain strain plot for these samples shown in Figure 12. As expected, the mechanical properties of these customised wereshown weaker samples entirely with optimum parameter A (200 W) plot for theseparts samples in than Figure 12. As produced expected, the mechanical properties of these customised customised parts were weaker than samples produced entirely with optimum parameter A (200 W) settings (Table 4), this is because approximately 1.7% of the SLM gauge length was produced using parts were weaker than samples produced entirely with optimum parameter A (200 W) settings settings (Table 4), this is because approximately 1.7% of the SLM gauge length was produced using parameter set C (150 W) marginally increasing porosity within the component across 10 layers (500 (Table 4), this is because approximately 1.7% of the SLM gauge length was produced using parameter parameter set C (150 W) marginally increasing porosity within the component across 10 layers (500 µm). When single parameters of set the A component samples, customised samples held set C (150 W) comparing marginally increasing porosity within across 10 layers (500 set µm).DWhen µm). When comparing single parameters of set A samples, customised samples set D held approximately 4% lower yield strength and 10% lower UTS. Customised sample sets E and F showed comparing single parameters of set A samples, customised samples set D held approximately 4% approximately 4% lower yield strength and 10% lower UTS. Customised sample sets E and F showed alower yield strength comparable to that of Customised sample set A, while the UTSF showed reduction of strength and 10% lower UTS. sets E and showed aaayield strength a yieldyield strength comparable to that of sample set sample A, while the UTS showed reduction of approximately 5.5%. comparable to that of sample set A, while the UTS showed a reduction of approximately 5.5%. approximately 5.5%. Table values for for customised customised specimen specimen sets. sets. Table 6. 6. Mean Mean mechanical mechanical property property values Table 6. Mean mechanical property values for customised specimen sets.

Specimen

D D D Property Property 0.2% Yield point (MPa)421 ± 421 0.2% Yield point (MPa) 12 ± 12 0.2% Yield point (MPa)512 421 ± 12 UTS (MPa) ± 7 ±7 UTS (MPa) 512 (MPa) 512 Fracture stressUTS (MPa) 509 ± 7 ±7 Fracture stress (MPa) 509 ± 7 Fracture stress (MPa) 509 ± 7 Specimen Specimen

Property

E EE 443 99 443± ± 443 ±±9 3 536 536 ± 3 536 529±±3 3 529 ± 3 529 ± 3

F F F 442 ± 12442 ± 12 442 ± 12533 ± 9 533 ± 9 533 ± 9 526 ± 9 526 ± 9 526 ± 9

Technologies 2017, 5, 9 Technologies 2017, 5, 9

9 of 12 9 of 12

Figure Figure 12. 12. Customised Customised specimen specimen stress-strain stress-strain plot. plot.

4. 4. Discussion Discussion The The results results obtained obtained for for the the porosity porosity of of the the samples samples in in Figure Figure 44 clearly clearly indicate indicate an an increase increase in in porosity and pore size when progressing from parameter set A to C. The increased porosity may have porosity and pore size when progressing from parameter set A to C. The increased porosity may have arisen arisen due due to to lack lack of of fusion fusion or or Rayleigh Rayleigh instability, instability, balling, balling, or or poor poor wetting wetting characteristics characteristics as as detailed detailed by Rombouts, M. et al. [12]. With regards to the microstructure, long elongated grains were observed by Rombouts, M. et al. [12]. With regards to the microstructure, long elongated grains were observed parallel parallel to to the the building building direction direction and and consistent consistent with with other other SLM SLM work work processing processing 316L 316L [7]. [7]. These These grains grains exhibit is is common in in most 300300 series stainless steels (316L included) [11,34]. The exhibitaustenitic austeniticbehaviour behaviourasas common most series stainless steels (316L included) [11,34]. production of austenite grains is highly temperature and cooling rate dependent, with minimal The production of austenite grains is highly temperature and cooling rate dependent, with minimal variation observed in in microstructures microstructuresproduced producedusing usingsamples samples sets and these observations variation observed sets A,A, B, B, and C, C, these observations are are consistent with other work on SLM processing 316L variable processing parameters consistent with other work on SLM processing 316L withwith variable processing parameters [7]. [7]. From From the the stress-strain stress-strain curve curve of of the the uniform uniform sample sample set set in in Figure Figure 7, 7, it it is is clear clear that that even even though though the energy density transferred to the sample was kept constant for samples A, B, and C, the energy density transferred to the sample was kept constant for samples A, B, and C, the the change change in parameters had an effect on the density of samples and subsequent mechanical properties. in parameters had an effect on the density of samples and subsequent mechanical properties. UTS UTS and and fracture fracture strength strength have have relatively relatively close close values, values, indicating indicating signs signs of of brittle brittle behaviour, behaviour, additionally additionally supported/explained supported/explainedby bythe thereduced reducednecking neckingat atthe thespecimen specimenfracture fracture point, point, results results consistent consistent with with work conducted by Riemer, A., et al. [7]. A notable effect was observed in the 0.2% yield stress, UTS, work conducted by Riemer, A., et al. [7]. A notable effect was observed in the 0.2% yield stress, UTS, and and fracture fracture stress, stress, with with aa 29% 29% decrease decrease in in yield, yield, 33% 33% decrease decrease in in UTS, UTS, and and 38% 38% decrease decrease in in fracture fracture when when progressing progressing from from parameter parameter set set A A to to C. C. There There is is aa high high variance variance in in the the measured measured maximum maximum strain strain results results with with an an average average standard standard deviation deviation of of 19%. 19%. However, However, there there isis still still aa clear clear trend trend of of decreasing max strain with decreasing laser power and increased exposure time, with a 49% decrease decreasing max strain with decreasing laser power and increased exposure time, with a 49% decrease in maximum strain strainfrom fromparameter parameterset setAAtotoC.C.This Thisdecrease decrease average maximum strain in average average maximum inin average maximum strain cancan be be attributed to the increase in porosity. According to ref. [35], decrease in porosity resulted in a attributed to the increase in porosity. According to ref. [35], decrease in porosity resulted in a significant significant improvement in Additionally, elongation. Additionally, the larger maximum strain— improvement in elongation. set A—havingset theA—having larger maximum strain—exhibits more of exhibits more of a ductile behaviour by absorbing more energy per unit volume (area under stressa ductile behaviour by absorbing more energy per unit volume (area under stress-strain curve), and strain curve), and conversely by having a significantly lower maximum strain, set C exhibits a more conversely by having a significantly lower maximum strain, set C exhibits a more brittle behaviour brittle behaviour with less unit of In volume absorbed. addition tofrom thesethe observations with less energy per unit of energy volumeper absorbed. addition to theseIn observations tensile test, from the tensile test, the hardness testing results from Table 5 shows a trend of decreasing hardness the hardness testing results from Table 5 shows a trend of decreasing hardness with decreased laser with decreased laser power and increased exposure time, with a 21% decrease progressing from power and increased exposure time, with a 21% decrease progressing from set A to C. Even with useset of A to C. Even with use of a constant energy density delivered to the SS 316L powder, variation in the a constant energy density delivered to the SS 316L powder, variation in the process parameters leads process parameters leads to changes in the material behaviour. to changes in the material behaviour. For the customised specimens, samples fractured at the designed failure location which were created through processing specific layers with “weaker” processing parameters as in set C, while maintaining set A properties throughout the remainder of the sample. This introduces a predictable

Technologies 2017, 5, 9

10 of 12

For the customised specimens, samples fractured at the designed failure location which were created through processing specific layers with “weaker” processing parameters as in set C, while maintaining set A properties throughout the remainder of the sample. This introduces a predictable failure point within samples, and prevents randomness of fracture across the gauge length of samples (as those seen in uniform parameter set samples, Figure 6). The added capability of designing fracture locations within samples results using “hybrid” processing parameters (sets D, E, F) caused samples to have approximately a 2% lower yield strength and 7% lower UTS than standard uniforms samples (processed set A). 5. Conclusions Whilst maintaining a consistent energy density, sample porosity varied as laser power and exposure time was modified. This change is linked to lack of fusion porosity or Rayleigh instabilities within the melt pool. It was observed that tensile samples produced with a uniform parameter set would fracture at random locations across its gauge length. Using a set of “hybrid” processing parameters, sample mechanical properties could be tailored such that specific fracture points within a sample could be designed into a component. Using two SLM processing parameters sets (hybrid parameters) to melt specific locations within a test piece, an additional 3% porosity across a segment making up approximately 1.7% of the sample’s gauge length was generated. This was sufficient to initiate consistent and repeatable fracture across this segment while only reducing the yield strength of the entire sample by approximately 2% compared to uniform single set parameter specimens (these produced un-customised, random fracture locations). This controlled in-built failure of components requires particular features to be “written” into the structure of a component through modification of SLM processing parameters at specific layers across a component. Controlled in-built failure of components and customisation of components’ mechanical performance is a feature that is difficult to achieve with conventional metal forming techniques. Such customised components can form part of a larger assembly, which are intentionally engineered to exhibit particular behaviours when under specific external loading conditions in order to protect other components within the system. Examples of parts include rupture discs used for pressure relief in pressure vessels; blow-off panels used in enclosures, vehicles, or buildings where overpressure may occur; and shear pins preventing mechanical overloads. Using the approach of “hybrid” processing, parameters, and porosity within samples may be graded/adjusted layer to layer. This customisation of mechanical properties will exert more refined control over specific part fracture points or general mechanical behaviour in order to further enhance the overall performance or capabilities of a component. As research momentum in this underdeveloped area progresses, additive manufacturing will be able to lay claim and present an additional benefit of this technology. Further exploitation of the unique layer-by-layer building principle of SLM can lead to customisation of components’ mechanical performance. Author Contributions: Andrei Ilie planned experimentation and undertook majority of testing, Haider Ali assisted with testing, analysis and manuscript preparation and Kamran Mumtaz supervised overall project. Conflicts of Interest: The authors declare no conflict of interest.

References 1. 2. 3. 4.

Childs, T.H.C.; Hauser, C.; Badrossamay, M. Mapping and Modelling Single Scan Track Formation in Direct Metal Selective Laser Melting. CIRP Ann. Manuf. Technol. 2004, 53, 191–194. [CrossRef] Yadroitsev, I.; Bertrand, P.; Smurov, I. Parametric analysis of the selective laser melting process. Appl. Surf. Sci. 2007, 253, 8064–8069. [CrossRef] Niendorf, T.; Leuders, S.; Riemer, A.; Richard, H.A.; Tröster, T.; Schwarze, D. Highly Anisotropic Steel Processed by Selective Laser Melting. Metall. Mater. Trans. B 2013, 44, 794–796. [CrossRef] Yadroitsev, I.; Thivillon, L.; Bertrand, P.; Smurov, I. Strategy of manufacturing components with designed internal structure by selective laser melting of metallic powder. Appl. Surf. Sci. 2007, 254, 980–983. [CrossRef]

Technologies 2017, 5, 9

5. 6. 7.

8. 9.

10. 11.

12. 13. 14. 15.

16. 17.

18.

19. 20.

21.

22.

23. 24. 25.

11 of 12

Fischer, P.; Romano, V.; Weber, H.P.; Karapatis, N.P.; Boillat, E.; Glardon, R. Sintering of commercially pure titanium powder with a Nd:YAG laser source. Acta Mater. 2003, 51, 1651–1662. [CrossRef] Kruth, J.P.; Levy, G.; Klocke, F.; Childs, T.H. Consolidation phenomena in laser and powder-bed based layered manufacturing. CIRP Ann. Manuf. Technol. 2007, 56, 730–759. [CrossRef] Riemer, A.; Leuders, S.; Thöne, M.; Richard, H.A.; Tröster, T.; Niendorf, T. On the fatigue crack growth behavior in 316L stainless steel manufactured by selective laser melting. Eng. Fract. Mech. 2014, 120, 15–25. [CrossRef] Yasa, E.; Kruth, J.P. Microstructural investigation of Selective Laser Melting 316L stainless steel parts exposed to laser re-melting. Procedia Eng. 2011, 19, 389–395. [CrossRef] Mertens, A.; Reginster, S.; Contrepois, Q.; Dormal, T.; Lemaire, O.; Lecomte-Beckers, J. Microstructures and mechanical properties of stainless steel AISI 316L processed by selective laser melting. Mater. Sci. Forum 2014, 783, 898–903. [CrossRef] Antony, K.; Arivazhagan, N.; Senthilkumaran, K. Numerical and experimental investigations on laser melting of stainless steel 316L metal powders. J. Manuf. Processes 2014, 16, 345–355. [CrossRef] Zhang, B.; Dembinski, L.; Coddet, C. The study of the laser parameters and environment variables effect on mechanical properties of high compact parts elaborated by selective laser melting 316L powder. Mater. Sci. Eng. A 2013, 584, 21–31. [CrossRef] Rombouts, M.; Kruth, J.P.; Froyen, L.; Mercelis, P. Fundamentals of Selective Laser Melting of alloyed steel powders. CIRP Ann. Manuf. Technol. 2006, 55, 187–192. [CrossRef] Simchi, A. Direct laser sintering of metal powders: Mechanism, kinetics and microstructural features. Mater. Sci. Eng. A 2006, 428, 148–158. [CrossRef] Roberts, I.A. Investigation of Residual Stresses in the Laser Melting of Metal Powders in Additive Layer Manufacturing; University of Wolverhampton: Wolverhampton, UK, 2012; p. 246. Knowles, C.R.; Becker, T.H.; Tait, R.B. The effect of heat treatment on the residual stress levels within direct metal laser sintered Ti-6Al-4V as measured using the hole-drilling strain gauge method. In Proceedings of the 13th international Rapid Product Development Association of South Africa (RAPDASA) Conference, Sun City, South Africa, 31 October–2 November 2012; pp. 1–10. Chatterjee, A.N.; Kumar, S.; Saha, P.; Mishra, P.K.; Choudhury, A.R. An experiment design approach to selective laser sintering of low carbon steel. J. Mater. Process. Technol. 2003, 136, 151–157. [CrossRef] Matsumoto, M.; Shiomi, M.; Osakada, K.; Abe, F. Finite element analysis of single layer forming on metallic powder bed in rapid prototyping by selective laser processing. Int. J. Mach. Tools Manuf. 2002, 42, 61–67. [CrossRef] Kruth, J.P.; Deckers, J.; Yasa, E.; Wauthlé, R. Assessing and comparing influencing factors of residual stresses in selective laser melting using a novel analysis method. Proc. Inst. Mech. Eng. Part B J. Eng. Manuf. 2012, 226, 980–991. [CrossRef] Joe Elambasseril, S.F.; Matthias, B.; Milan, B. Influence of Process Parameters on Selective Laser Melting of Ti 6Al-4V Components; RMIT University: Ho Chi Minh City, Vietnam, 2012. Papadakis, L.; Loizou, A.; Risse, J.; Bremen, S. A thermo-mechanical modeling reduction approach for calculating shape distortion in SLM manufacturing for aero engine components. In Proceedings of the 6th International Conference on Advanced Research in Virtual and Rapid Prototyping, Leiria, Portugal, 1–5 October 2013. Casavola, C.; Campanelli, S.L.; Pappalettere, C. Experimental analysis of residual stresses in the selective laser melting process. In Proceedings of the XIth International Congress and Exposition, Orlando, FL, USA, 2–5 June 2008. Thöne, M.; Leuders, S.; Riemer, A.; Tröster, T.; Richard, H.A. Influence of heat-treatment on selective laser melting products–eg Ti6Al4V. In Proceedings of the Solid Freeform Fabrication Symposium SFF, Austin, TX, USA, 6–8 August 2012. Shiomi, M.; Osakada, K.; Nakamura, K.; Yamashita, T.; Abe, F. Residual Stress within Metallic Model Made by Selective Laser Melting Process. CIRP Ann. Manuf. Technol. 2004, 53, 195–198. [CrossRef] ASTM International. Additive Manufacturing Technology Standards; ASTM International: West Conshohocken, PA, USA, 2014. Wohlers Associates, Inc. Wohlers Report 2008: State of the Industry; Annual Worldwide Progress Report; Wohlers Associates, Inc.: Fort Collins, CO, USA, 1996.

Technologies 2017, 5, 9

26. 27. 28. 29. 30. 31. 32. 33. 34. 35.

12 of 12

Gu, D.; Shen, Y. Balling phenomena in direct laser sintering of stainless steel powder: Metallurgical mechanisms and control methods. Mater. Des. 2009, 30, 2903–2910. [CrossRef] Yadroitsev, I.; Gusarov, A.; Yadroitsava, I.; Smurov, I. Single track formation in selective laser melting of metal powders. J. Mater. Process. Technol. 2010, 210, 1624–1631. [CrossRef] Hanzl, P.; Zetek, M.; Bakša, T.; Kroupa, T. The Influence of Processing Parameters on the Mechanical Properties of SLM Parts. Procedia Eng. 2015, 100, 1405–1413. [CrossRef] Vrancken, B.; Thijs, L.; Kruth, J.P.; Van Humbeeck, J. Heat treatment of Ti6Al4V produced by Selective Laser Melting: Microstructure and mechanical properties. J. Alloys Compd. 2012, 541, 177–185. [CrossRef] ISO 6507-1:2005, Metallic Materials—Vickers Hardness Test—Test Method. Available online: http://www. iso.org/iso/catalogue_detail.htm?csnumber=37746 (accessed on 20 February 2017). ASTM International. Standard Test Methods for Determining Area Percentage Porosity in Thermal Sprayed Coatings; E2109-01(2007); ASTM International: West Conshohocken, PA, USA, 2007. British Standards Institution (BSI). Method for Statistically Estimating the Volume Fraction of Phases and Constituents by Systematic Manual Point Counting with a Grid; BS 7590:1992; BSI: London, UK, 1992. ASTM International. Standard Test Methods for Tension Testing of Metallic Materials; E8/E8M-13a; ASTM International: West Conshohocken, PA, USA, 2013. Reed-Hill, R.E.; Abbaschian, R. Physical Metallurgy Principles; PWS-Kent Publisher: Boston, MA, USA, 1992. Qiu, C.; Adkins, N.J.E.; Attallah, M.M. Selective laser melting of Invar 36: Microstructure and properties. Acta Mater. 2016, 103, 382–395. [CrossRef] © 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).